Methods for normalization of hippocampal volumes measured with MR.

PURPOSE To investigate the use of six cerebral measures as correlates for hippocampal volumes and, therefore, to enable normalized absolute hippocampal volumes to be calculated via two correction processes. METHODS Hippocampal volumes and six cerebral measures were estimated from MR data in 20 control subjects. Three of these measures (the cranial volume, the cerebral volume, and the midsagittal cranial area) were then applied to a group of 32 control subjects, and regression analysis was performed to investigate the linear relationship between hippocampal volume and each measure. Division of hippocampal volume by cerebral measure and correction via a covariance calculation enabled corrected absolute hippocampal volumes to be determined for 32 control subjects and 23 patients with temporal lobe epilepsy. RESULTS Correction processes reduced the variance in absolute hippocampal volumes in control subjects and enabled abnormally small absolute volumes to be defined. Of 11 patients with unilateral volume ratio abnormalities, 8 had unilateral abnormally small absolute hippocampal volumes. Of 12 patients with normal volume ratios, 4 had bilateral abnormally small absolute hippocampal volumes. CONCLUSION Correction processes can define absolute hippocampal volumes for correlation studies and may enable identification of unsuspected bilateral hippocampal volume loss.